The present invention relates generally to a system and method for data communication and storage. More particularly, the present invention relates to both lossless and lossy compression techniques to concentrate relevant signal information for data communication and storage.
There are a variety of techniques for the compression of data which may be stored digitally or by analog techniques. This data represents audio, visual or other information for which there are numerous and different constraints to their compression.
Computer data files are, with few exceptions, stored digitally. The inherent advantages of digital communication and storage techniques are primarily due to the fact that information which is transmitted and stored in a binary form is much less susceptible to corruption due to noise or distortion than conventional analog systems. In addition, the conversion of analog signals to a digital form enables the user to employ noise reduction techniques and advanced signal processing algorithms which cannot typically be conducted on conventional analog signals. Moreover, digital communication and storage can also provide exact reproduction of the system output signals.
This is important because computer files must retain all of their information during storage or transmission, as an error of even one bit can totally corrupt the file. Corruption of a computer file can result in either failure of the entire computer system, or in the case of corruption of a non-operating file, such as a word processing file, inability to recreate the exact copy of the original file. Accordingly, the storage or transmission of computer data files or the like require highly reliable systems which maintain file integrity. Moreover, any compression of a computer file must be lossless, meaning not result in the loss of any data and provide exact reproduction of the uncompressed data.
Unfortunately, digital transmission and storage techniques disadvantageously require much wider frequency bandwidth. This is particularly true with respect to video information and modern multimedia systems which require the processing and storage of high volumes of data. Moreover, the transmission of video must often be accomplished in real time wherein the video information is transmitted at the same rate or faster than video playback.
To illustrate the complexity of the problem, each channel of the Common Interface Format (CIF) resolution standard for video systems requires 352 pixels per line and 288 lines per frame. In addition, CIF requires 2 chrominace channels with half resolution of 176 pixels per line, 146 lines per frame and 8 bits per pixel. With full motion video requiring approximately 30 frames per second transmission rate, video transmission requires approximately 36 Mbit per second. By means of an example, using a 56.6 k bits per second modem, it would take almost 11 minutes to transmit 1 second of video data over a telephone line. As a separate example, a CD-ROM having a capacity of 650 megabytes can only store approximately 18 seconds of uncompressed (CIF) video data.
Present systems do not provide adequate transmission rate of video signals over available communication channels. Presently, the most popular and inexpensive means for transmission of digital data is through the public telephone network. Unfortunately, the public telephone network was designed to transmit analog signals in a voice frequency range which is significantly lower than the frequency range required for most present day communication systems including digital data, voice and video communication.
To transmit digital information over the public telephone network, one typically uses a modem. Digital information is converted to an analog form. The modem filters the digital signal by shifting the signal and frequency to form a band limited signal. The modem then modulates that signal within the bandwidth of the communications channel which is typically between 300 Hz and 3500 Hz. Present modems employ quadrature modulation to increase the transmission rate of the digital information. Unfortunately, quadrature modulation has only increased present modem speed to approximately 56,600 bits per second. Accordingly, for practical digital transmission or storage, it is necessary to reduce the amounts of data to be transmitted or stored by either eliminating redundant information or by reducing the quality of the information.
As opposed to lossless compression techniques, data that is ultimately observed by the human senses can often be compressed with the loss of some information without any discernable alteration, as the human senses have limited capabilities in perception. Accordingly, audio and visual information is often compressed by lossy methods, for which there is a loss of information, since it is only necessary to recreate (decompress) a signal to the degree required for a subjective quality level rather then to perfectly recreate the signal.
For example, U.S. Pat. No. 5,819,215 issued to Dobson, U.S. Pat. No. 5,812,915 issued to Zhang and U.S. Pat. No. 5,845,243 issued to Smart each disclose lossy compression methods. U.S. Pat. Nos. 5,819,215 and 5,845,243 teach a wavelet based compression method having an adaptive bit rate control. An analog signal is digitally sampled at a desired rate and samples are transformed into the wavelet domain to form wavelet coefficients. The data is then compressed by reducing the total number of coefficients to be saved.
U.S. Pat. No. 5,812,195 teaches compressing video by using a prediction video signal to predict image blocks of a video image. An error measurement is obtained from comparing the predicted image to the actual image.
U.S. Pat. No. 5,426,665 issued to Cleverley et al. discloses a compression technique for spread spectrum communication systems. For spread spectrum communication systems or radar pulse compression systems, data is mixed with a pseudo random code with the frequency alternated during radio transmission. In a receiver, the process is reversed with the received signal down converted and then decoded to extract the data.
U.S. Pat. No. 5,184,229 issued to Saito et al. discloses a data compression system utilizing Huffman coded data.
Meanwhile, U.S. Pat. No. 5,818,870 issued to Yaguchi discloses transmitting an analog signal through a digital communication device. The amplitude to the signal is modulated to provide analog information, while the frequency is maintained at the normal rate of the digital communication device.
U.S. Pat. No. 5,661,718 issued to Bremer et al. discloses the simultaneous transmission of analog and digital communications. A sampled analog signal can be quantisized and represented in digital form. The analog signal which is then converted to digital form is then converted to amplitude quantisized pulse amplitude modulated format, such as conventional PCM.
Unfortunately, in such prior art compression techniques, the data transmission rates are still too slow for many practical applications. In addition, amplitude modulation of a transmitted signal often suffers from signal-to-noise ratio (S/N) problems. Moreover, it would be advantageous to provide a compression system and apparatus which did not result in the loss of any information during storage or transmission.
Briefly, in accordance with the invention, I provide an improved apparatus and method for compressing digital data by converting the digital data to an analog signal. Generally, it is thought that information, stored in digital or in analog format, can be more quickly or more efficiently transmitted by digital processing than by known analog methods. Contrary to past teachings, I have discovered that digital information can be more quickly or efficiently transmitted if first translated into an analog format.
Digital data is comprised of xe2x80x9cbitsxe2x80x9d in the form of xe2x80x9conesxe2x80x9d and xe2x80x9czerosxe2x80x9d. For example, applying traditional digital techniques, a signal transmitted at 3000 Hz provides the opportunity to transmit 3000 individual bits in one second. These individual ones and/or zeros are then complied in discreet packages called xe2x80x9cwordsxe2x80x9d. For example, typically an 8 bit processor compiles and processes 8 bit words, and converts those words into usable information.
My invention provides for converting the digital data into an analog signal for storage or transmission. The digital data is first separated into digital words. Each digital word, including for example 8 bits, is then assigned a preselected frequency assignment according to a predetermined conversion table. An analog signal is then created having an alternating frequency with the frequency alternating every predetermined number of cycles. The alternating frequencies of the analog signal are assigned by sequentially arranging the preselected frequency assignments corresponding to the digital words. More particularly, the analog signal is modulated in frequency every predetermined number of cycles with each frequency corresponding to a preselected frequency assignment and each frequency assignment defining a digital word.
This method should not be confused with standard frequency modulation (FM), which it is not. A typical FM signal is modulated from a single defined frequency assignment according to the information, such as sound, being stored or transmitted. In contrast, according to the present invention, the analog signal is not modulated from a single frequency. Instead, like Code Diversion Multiple Access (CDMA), the alternating frequency (hereinafter xe2x80x9cAFxe2x80x9d) analog signal includes a plurality of alternating frequency assignments. However, unlike CDMA it is the particular frequency assignment which provides information corresponding to the original digital data.
For example only, according to an embodiment of the present invention, a system having a maximum frequency rate of 4000 Hz is formatted to include eight (8) frequency assignments between 2250 Hz and 4000 Hz, with each frequency assignment having a bandwidth of 250 Hz. Digital data is separated into 3 bit words and each word is assigned a particular frequency assignment, for example, according to the following conversion table.
A digital stream, such as 010101000110 . . . , is broken into individual 3 bit words, such as 010, 101, 000 and 110. An AF analog signal is then generated with the analog signal modulated in frequency every predetermined number of cycles with each frequency corresponding to preselected frequency assignments and each frequency assignment defining a digital word according to the conversion table. For example, an AF analog signal alternating every one (1) cycle based on the above digital stream and above conversion table would include a first cycle at 2750 Hz (corresponding to xe2x80x9c010xe2x80x9d), a second cycle at 3500 Hz (corresponding to xe2x80x9c101xe2x80x9d), a third cycle at 2250 Hz (corresponding to xe2x80x9c000xe2x80x9d), and a fourth cycle at 3750 Hz (corresponding to xe2x80x9c110xe2x80x9d). In this manner twelve (12) digital bits can be transmitted in four cycles of the analog signal at an average frequency of approximately 3000 Hz, instead of transmitting four (4) digital bits in four cycles for an analog signal having a frequency of 4000 Hz.
The AF analog signal is then transmitted or stored by means known to those skilled in the art. For example, the AF analog signal may be stored by magnetic tape, or transmitted such as by wire, fiberoptics, RF transmission, or the like.
To decompress the AF analog signal and reproduce the original digital data, the AF analog signal is processed with the alternating frequencies of the analog signal being measured. The alternating frequencies are, in turn, converted back into-digital words according to the predetermined conversion table, and the digital words are sequentially arranged to produce a digital data stream. The digital data stream is a reproduction of original data that was first converted into an AF analog signal.
In preferred embodiments, digital data is further compressed by reducing the bandwidth between frequency assignments, by altering the predetermined number of cycles before the analog signal is alternated in frequency, and by increasing the number of digital bits which correspond to each word. For example, in a preferred embodiment, a much greater data transmission rate is achieved by alternating the frequency assignments every half signal cycle instead of every signal cycle. In addition, with reference to the above example, a greater data transmission rate can be achieved by reducing the frequency bandwidth between frequency assignments from 250 Hz to 125 Hz, or even 50 Hz, thereby raising the average frequency rate.
In an additional preferred embodiment, the conversion tables are created such that the more frequently used digital words correspond to frequency assignments having shorter wavelengths. For example, the digital word used most often in a particular system, or during a particular application, is assigned a frequency assignment having the highest frequency. The predominance of digital words having shorter wavelengths will further increase the data transmission rate. In an additional preferred embodiment, the digital data is sampled, or analyzed in its entirety, by the first convertor processor. The first converter processor then selects, or creates, a conversion table wherein the predominantly used digital words are assigned to frequency assignments having the greatest frequency. In this manner, the AF analog signal will have a comparatively higher average frequency for the system, and comparatively greater data transmission rate.
As would be understood by those skilled in the art, the present invention provides for encryption of data. Further encryption of the AF analog signal can also be simply accomplished by using proprietary conversion tables. Moreover, as long as the same conversion table is used for encoding and decoding, the conversion tables can be changed or modified to provide a high degree of protection which can be accomplished by using multiple conversion tables.
Where the measurement means for measuring the frequency assignments of the analog signal is not very precise, it is preferred that the predetermined number of cycles is increased before the analog signal is alternated in frequency. For example, the predetermined number of cycles may be increased to 10, or even 100 cycles, to permit the frequency measurement means to make a substantial number of frequency measurements of the analog signal to provide an extremely accurate and verified determination of an alternating frequency before measuring the frequency corresponding to the next digital word.
Though the present invention has been explained with reference to the compression of digital data, analog data may also by compressed by application of the present invention. Analog data is first converted into digital data by means know to those skilled in the art, for example by an analog-to-digital converter. The digital data is then converted into an AF analog signal having alternating frequencies, as explained above. To decompress the AF analog signal and convert the signal back into the analog data""s original format, the AF analog signal is first converted into digital data, also by means explained above. The digital data is then processed through a digital-to-analog converter known to those skilled in the art, to reproduce the original analog data.
It is thus an object of the present invention to provide new and improved apparatus and methods for compressing analog and digital data.
It is another object of the present invention to provide lossless apparatus and methods for transmitting and storing compressed analog and digital data.
It is still another object of the present invention to provide apparatus and methods providing for the transmission of data over bandlimited transmission medium such as standard public telephone lines.
It is an additional object of the present invention to provide full-motion video communication over bandlimited transmission medium such as standard analog telephone lines.